The risk for intracranial hemorrhage is greater with alteplase therapy, especially in patients with high blood pressure [11-14]. Streptokinase has therefore been suggested as the agent of choice in patients with elevated blood pressure [15]. Unfortunately, many of the data on the relation between blood pressure and outcomes after thrombolysis have come from studies with small numbers of patients [8, 16, 17].

Our objective was to assess the influence of previous hypertension and blood pressure at study entry on the outcomes of death, any stroke, intracranial hemorrhage, and death plus disabling stroke using data from the GUSTO-I (Global Utilization of Streptokinase and t-PA for Occluded Coronary Arteries) trial of 41 021 patients. We reviewed data from hypertensive patients who were at low risk for death from cardiac causes to assess relative risks for stroke and other outcomes. We also reviewed the influence of blood pressure at the time of study entry on outcomes with different thrombolytic agents.

Methods

Top Methods Results Discussion Author & Article Info References

Patients

In the GUSTO-I trial, 41 021 patients with acute myocardial infarction and ST-segment elevation who presented within 6 hours of symptom onset were randomly assigned to one of four thrombolytic treatment regimens [18]. Patients were excluded if they had previously had stroke (or known cerebrovascular pathology, such as a neoplasm, an aneurysm, or intracranial surgery), had active bleeding or had recently had bleeding, had recently had trauma or major surgery, had noncompressible vascular punctures, or had previously received streptokinase or anistreplase. Hypertension (in the past or at the time of presentation) was not an exclusion criterion. However, if a patient's initial systolic blood pressure was 200 mm Hg or more (or the diastolic blood pressure was 100 mm Hg) at the time of randomization, the following statement was read to the investigator: "We have been advised to let you know that with a blood pressure this high there is an increased risk of intracranial bleeding. It would be advisable to bring the blood pressure down before starting thrombolytic therapy." The final decision about whether to enroll the patient was left to the investigator. To estimate whether patients enrolled in GUSTO-I who had very elevated blood pressure were a highly selected group, data from the GUSTO-I myocardial infarction registry, a randomly selected group of 92 hospitals (primarily in North America) that collected information on all patients who had had acute myocardial infarction, were analyzed.

Eligible patients were randomly assigned to receive one of four regimens: 1) streptokinase, 1.5 million U, and subcutaneous heparin, 12 500 U twice daily, beginning 4 hours after initiation of thrombolysis; 2) streptokinase, 1.5 million U, and intravenous heparin, a 5000-U bolus and 1000 U/h-infusion with dose adjustment to maintain an activated partial thromboplastin time of 60 to 85 seconds; 3) accelerated-dose alteplase, given as a 15-mg bolus, 0.75 mg/kg of body weight (to a maximum of 50 mg) over 30 minutes, and 0.5 mg/kg [to a maximum of 35 mg] over the next hour, and intravenous heparin, a 5000-U bolus and 1000 U/h-infusion with dose adjustment to maintain an activated partial thromboplastin time of 60 to 85 seconds; or 4) standard-dose alteplase, 1.0 mg/kg over 60 minutes (to a maximum of 90 mg) with 10% as a bolus, and streptokinase, 1.0 million U, with intravenous heparin, a 5000-U bolus and 1000 U/h-infusion with dose adjustment to maintain an activated partial thromboplastin time of 60 to 85 seconds. Subcutaneous heparin treatment was administered for 7 days or until hospital discharge; intravenous heparin was given for at least 48 hours. Chewable aspirin was given as soon as possible and then daily for at least 1 year. Intravenous ß-block-ade, unless contraindicated, was given as soon as possible and was followed with daily oral therapy. All other medications were administered at the discretion of the attending physician.

End Points

The primary end point was death from any cause within 30 days after randomization. Secondary end points included in-hospital stroke and primary intracranial hemorrhage. The composite of death and nonfatal, disabling stroke was intended to provide a single measure of outcome for comparisons of thrombolytic strategies that can reduce the risk for death from cardiac causes while increasing the risk for hemorrhagic stroke.

In the GUSTO-I protocol, stroke was defined as an acute, new neurologic deficit that resulted in death or lasted more than 24 hours. All cases of suspected stroke were reviewed by an independent committee, the members of which were blinded to treatment assignments. Computed tomography, magnetic resonance imaging, or autopsy was done in 93% of cases to enable classification of strokes. A complete description of the identification, categorization, and incidence of stroke in the GUSTO-I trial has been published [19].

Bleeding complications were classified as severe or life-threatening if they were intracerebral or if they caused substantial hemodynamic compromise that required intervention. Moderate bleeding was defined as bleeding that required transfusion.

Measurement of Blood Pressure

Two blood pressure measurements were recorded on each case report form: one taken at the time of presentation at the hospital and another taken just before randomization.

Statistical Analysis

For descriptive purposes, measurements of systolic and diastolic blood pressures at study entry were divided into categories. Measurements of systolic blood pressure at entry were divided into four strata: 100 to 124 mm Hg, 125 to 149 mm Hg, 150 to 174 mm Hg, and 175 mm Hg or more. Measurements of diastolic blood pressure at entry were divided into four strata: 70 to 79 mm Hg, 80 to 89 mm Hg, 90 to 99 mm Hg, and 100 mm Hg or more. Baseline characteristics are given as percentages for discrete variables and as medians (25th, 75th percentiles) for continuous characteristics. Continuous systolic and diastolic blood pressure variables were used for reporting P values.

Although analysis of the relation between blood pressure at entry and outcome was prospectively planned, our results should be interpreted in light of the retrospective nature of some of the specific analyses and the multiple comparisons done. Any results of borderline statistical significance should be considered exploratory and for the generation of hypotheses only.

The statistical importance of blood pressure at study entry with respect to clinical outcomes was computed by using logistic regression modeling techniques. Spline transformations [20, 21] were used for systolic and diastolic blood pressures in evaluating the assumption of a linear relation between these continuous variables and the logit of each outcome. In the few cases in which the relation was nonlinear, spline functions of the blood pressure measurements were used in the modeling process.

Both systolic blood pressure at study entry and a history of hypertension have been shown to be important predictors of death within 30 days in a multivariable model [22]. Diastolic blood pressure was added to this model to determine whether it would provide additional important prognostic information. Similarly, these three factors were added to a predictive model for intracranial hemorrhage that has been described elsewhere [19]. Because no model has been developed for the outcome of death plus nonfatal, disabling stroke, all variables found to be significant in either the mortality model or the stroke model were combined. The relative importance of the three factors was then evaluated within this new model.

We also assessed the differential effect of blood pressure on outcomes with respect to location of infarction and thrombolytic treatment assignment. For each blood pressure measurement and outcome, a univariable logistic model was first developed. The additive effect of infarction location or treatment assignment was evaluated with the addition of each factor separately to the univariable model. The differential effect of infarction location or treatment assignment was determined by testing the inclusion of the interaction factor for the blood pressure of interest, after adjustment for the two main effects.

Results

Top Methods Results Discussion Author & Article Info References

History of Hypertension

A history of hypertension was present in 15 544 patients (38.1%). The baseline characteristics of this group and those of the other 25 309 patients (62.1%) are shown in Table 1. Patients with previous hypertension were a higher-risk group: They were more likely to be older, female, diabetic, and in a worse Killip class and, as a group, they had more previous infarctions.

Patients with a history of hypertension had significantly higher rates of total stroke and combined death and disabling stroke Table 2, as well as higher rates of death [22] and intracranial hemorrhage [19], as described elsewhere. Patients with a history of hypertension also fared worse with regard to other clinical events, with the exception of the use of angioplasty.

Systolic Blood Pressure at Study Entry

The baseline characteristics of patients with various systolic blood pressures at entry are shown in Table 3. Measurements of systolic blood pressure at study entry were available for 40 214 patients (98%); the range was 0 to 270 mm Hg, and the median was 130 mm Hg (25th, 75th percentiles: 112 mm Hg, 144 mm Hg). Only 133 patients (0.33%) had a systolic blood pressure of 200 mm Hg or more at entry. Patients with higher systolic blood pressures had baseline characteristics associated with a higher baseline risk for death: They were more likely to be older, to have a higher heart rate, to be female, and to be diabetic; they were less likely to be smokers; and they were more likely to present with anterior infarction. As a group, they had the longest time to treatment.

Figure 1 shows the predicted probabilities of death within 30 days; intracranial hemorrhage; and death plus nonfatal, disabling stroke as functions of systolic blood pressure at study entry. More thorough analyses of the risk factors (other than hypertension) for death [22] and intracranial hemorrhage [19] have been reported elsewhere. The mortality rate was highest in the patients with the lowest blood pressures, and no substantial change was seen in the probability of death as systolic blood pressure at entry increased from 120 mm Hg. The rates of total stroke and intracranial hemorrhage increased as systolic pressure increased (P < 0.001) (Figure 1); the rate of intracranial hemorrhage was 1.71% for patients with a systolic blood pressure of 175 mm Hg or more at entry. The rate of death plus disabling stroke decreased significantly from the lowest systolic blood pressures to systolic blood pressures of approximately 120 mm Hg. At pressures greater than 120 mm Hg, the risk for this combined event was not increased, although there was a trend toward increasing risk with increasing pressure for very high systolic pressures (Figure 1). The rate of death plus disabling stroke was also highest among patients whose systolic pressures were 175 mm Hg or more (7.8%).

View larger version (23K): [in this window] [in a new window]   Figure 1. Predicted probabilities (solid lines) with 95% Cls (dashed lines) for death within 30 days (top), intracranial hemorrhage (middle), and death plus disabling stroke within 30 days (bottom) by regression analysis as functions of systolic blood pressure at study entry. P < 0.001 for the effect of systolic blood pressure for each end point.

The incidence of bleeding was not strongly related to systolic blood pressure at entry Table 4, especially when procedural bleeding (related to cardiac catheterization, bypass surgery, angioplasty, or intraaortic balloon pumping) was excluded. Rates of severe bleeding were 0.7%, 0.8%, 0.7%, and 0.9% for each of the four blood pressure groups, respectively. Patients with systolic blood pressures of 175 mm Hg or more had more recurrent ischemia and reinfarction.

Diastolic Blood Pressure at Study Entry

Measurements of diastolic blood pressure at study entry were available for 39826 patients (97%) and ranged from 0 mm Hg to 160 mm Hg; the median was 80 mm Hg (70 mm Hg, 90 mm Hg). Table 5 shows the baseline characteristics of patients for each stratum of diastolic blood pressure at study entry. Patients with lower diastolic blood pressures were older, weighed less, and were more often female. Time to treatment was similar among the groups. Table 6 shows clinical outcomes for each stratum of diastolic blood pressure at entry. No significant change in mortality rate was seen with diastolic pressures greater than 100 mm Hg. Rates of stroke and intracranial hemorrhage did increase somewhat as diastolic blood pressure increased. Patients with lower diastolic blood pressures fared worse with regard to each outcome with the exceptions of recurrent ischemia and reinfarction, which were similar among the groups.

Infarction Location and Elevated Arterial Pressure

A systolic blood pressure of 175 mm Hg or more at entry was present in 1229 patients, of whom 661 (54%) had an inferior infarction and 518 (42%) had an anterior infarction. As in the overall study, the mortality rate tended to be lower in hypertensive patients who had an inferior infarction than in hypertensive patients who had an anterior infarction (6.06% compared with 8.22%; P = 0.151). No difference was seen in rates of stroke (3.29% compared with 3.33%) or intracranial hemorrhage (1.66% compared with 1.74%). The rate of death plus disabling stroke also tended to be lower in patients who had an inferior infarction (7.87% compared with 10.23%; P = 0.157). The effect of systolic blood pressure on mortality; stroke or intracranial hemorrhage; or death plus nonfatal, disabling stroke did not differ significantly for patients with an inferior infarction and patients with an anterior infarction.

In a subgroup of hypertensive patients (systolic blood pressure 175 mm Hg) with low risk for death from cardiac causes (Killip class I and no previous infarction [n = 899]; 2% of the overall population), the overall mortality rate was low (4.33%) and the rates of total stroke (2.81%) and intracranial hemorrhage (1.29%) were high. The incidence of death plus disabling stroke was 6.20%.

In the 501 patients in this low-risk group who had an inferior infarction, the mortality rate was 4.39%; the rates of total stroke and intracranial hemorrhage were 2.59% and 1.39%, respectively; and the incidence rate of the combined end point was 5.78%. In the 353 patients in the low-risk group who had an anterior infarction, the mortality rate was 4.25%; the rates for stroke and intracranial hemorrhage were 3.07% and 1.12%, respectively; and the incidence rate of the combined end point was 6.69%.

Relation of Thrombolytic Agents to Systolic Blood Pressure at Study Entry

Figure 2 shows the predicted probabilities of death; intracranial hemorrhage; and death plus nonfatal, disabling stroke as functions of systolic blood pressure at entry and assignment to accelerated alteplase or streptokinase therapy (combined streptokinase arms). No statistical interaction was seen between the effect of treatment assignment (alteplase or streptokinase) and blood pressure on rates of death within 30 days of randomization (P > 0.2) or death plus disabling stroke (P > 0.2), indicating that there was no evidence of a differential effect of alteplase compared with streptokinase for different strata of arterial blood pressure at study entry.

View larger version (27K): [in this window] [in a new window]   Figure 2. Predicted probabilities with 95% Cls for death within 30 days (top), intracranial hemorrhage (middle), and death plus disabling stroke within 30 days (bottom) by regression analysis as functions of systolic blood pressure at study entry and treatment assignment; accelerated alteplase (solid lines) and the streptokinase monotherapies combined (dashed lines). For a comparison of the effect of accelerated alteplase with that of streptokinase after adjustment for systolic blood pressure, P = 0.002, 0.083, and 0.01, respectively.

The rates of stroke and intracranial hemorrhage increased as systolic pressure increased (Table 7); this increase was seen with both alteplase and streptokinase. Because the absolute risk for intracranial hemorrhage was higher with alteplase, the increase in the rate of intracranial hemorrhage as a function of higher blood pressure was greater in patients treated with accelerated alteplase. As systolic blood pressure increased from the lowest to the highest stratum, the overall risk for stroke increased from 1.1% to 3.4% and the risk for intracranial hemorrhage increased from 0.45% to 1.7%. The increases in the rate of stroke with increasing blood pressure were similar with the two thrombolytic agents Table 7: A 3.05-fold increase (from 0.97% to 2.96%) was seen among patients treated with streptokinase and a 3.09-fold increase (from 1.16% to 3.59%) was seen among patients treated with alteplase. Among patients with systolic blood pressures of 175 mm Hg or more, the absolute rate of nonfatal disabling stroke was 0.50% greater in the alteplase group than in the streptokinase group, whereas the rate of intracranial hemorrhage was 0.81% greater in the alteplase group. Although the overall rate of intracranial hemorrhage was higher with accelerated alteplase than with streptokinase, accelerated alteplase was associated with a lower combined rate of death plus disabling stroke, even among patients with a systolic blood pressure greater than 175 mm Hg at entry (4.9% compared with 8.9%; P = 0.031).

Discussion

Top Methods Results Discussion Author & Article Info References

History of Hypertension

Data were available for more than 15 000 patients with a history of hypertension. As in previous studies [4, 8-11], history of hypertension was a significant predictor of death, stroke, and bleeding. However, a history of hypertension was only a small determinant of risk for these events. Accordingly, it does not appear that a history of hypertension should influence the decision to administer thrombolytic therapy.

Blood Pressure at Study Entry

Except for systolic blood pressures of less than 120 mm Hg, with which the risk for death was greater than that seen with higher blood pressure strata, no significant relation between different systolic blood pressures and mortality was seen.

The risk for intracranial hemorrhage increased as systolic blood pressure increased. There was no threshold for this effect; however, at systolic pressures of about 170 mm Hg, the probability curve for intracranial hemorrhage became progressively steeper (Figure 1). The rate of intracranial hemorrhage seen in the patients who had systolic blood pressures of 175 mm Hg or more was double that of the rest of the cohort. This is consistent with data from previous studies [4-11]. In the review by Simoons and colleagues [8], blood pressure greater than 170 mm Hg was identified as one of four major risk factors for intracranial hemorrhage, with a risk ratio of 2.0.

Treatment of Patients with Elevated Blood Pressure and Low Risk for Death from Cardiac Causes

To estimate the risks and benefits of thrombolytic therapy among hypertensive patients at low risk for death from cardiac causes, we assumed that thrombolytic therapy reduced the rate of death from cardiac causes by approximately 30%. This estimate was derived from the overview of the Fibrinolytic Therapy Trialists' Collaborative Group [4], in which a 25% to 30% reduction in the overall mortality rate was seen among patients with ST-segment elevation who were enrolled within 6 hours of symptom onset. The increase in the rate of death attributable to excess intracranial hemorrhage caused by thrombolytic therapy was equivalent to 5% to 10% of the overall reduction in mortality in the patients who received thrombolytic therapy. Thus, the estimated relative reduction in the rate of death from cardiac causes due to thrombolysis is slightly greater than the estimated reduction in the overall mortality rate. Assuming that the absolute rate of death from cardiac causes without thrombolytic therapy among the low-risk hypertensive patients is 4.4%, a 1.3% absolute reduction in deaths from cardiac causes would occur. The rate of intracranial hemorrhage was 1.29% among "low-risk" patients with systolic blood pressures of 175 mm Hg or more. Therefore, assuming that the increase in the rate of stroke due to thrombolytic therapy is approximately equivalent to the rate of intracranial hemorrhage, 13 lives per 1000 patients treated with thrombolysis are being saved at the risk of 13 additional strokes.

Among these low-risk hypertensive patients, infarction site did not substantially affect the incidence of stroke; even patients who had an anterior infarction, were in Killip class I, and had had no previous infarction had a low mortality rate and a relatively high risk for stroke. Data from the small subgroups of patients who had hypertension and no heart failure (who comprised only 2.2% of the overall study sample) and patients who had an anterior infarction, hypertension, and no heart failure (only 0.9% of the overall study sample) should be interpreted with caution. The risk-to-benefit ratio of intravenous thrombolysis in patients who have low risk for death from cardiac causes but a systolic blood pressure of 175 mm Hg or more should be studied further.

Our data do not allow us to determine the correct management of patients with systolic blood pressures of 175 mm Hg or more. Some have advised decreasing the blood pressure as soon as possible to less than 170 mm Hg [11], but the effect of this maneuver on the incidence of death and stroke has not been investigated. Others suggest that patients in this group should have acute percutaneous transluminal coronary angioplasty because they are at high risk for stroke [23, 24]; however, few data on angioplasty are available for these patients, who would also be at high risk for bleeding complications of invasive procedures. The patients at low risk for death from cardiac causes present a particular problem because the risks of an invasive procedure may outweigh the procedure's benefits.

Comparison of Thrombolytic Agents: Relation to Blood Pressure at Study Entry

Compared with patients who received streptokinase therapy, those who received accelerated alteplase therapy had an increased rate of intracranial hemorrhage at all systolic blood pressure strata at study entry. The difference in risk for stroke between patients receiving the agents was magnified at systolic pressures greater than 175 mm Hg because the absolute risk for stroke was higher in this group, regardless of treatment assignment. This finding is in keeping with results of previous studies that compared alteplase with streptokinase. The rate of intracranial hemorrhage among patients who received streptokinase also increased as systolic blood pressure increased (substantially more so at systolic pressures of 175 mm Hg or more) but to a lesser extent than among patients who received accelerated alteplase therapy.

Despite the increased risk for stroke, the rate of the combined end point of death plus disabling stroke among patients with systolic blood pressures of 175 mm Hg or more was lower among those who received alteplase than among patients with this blood pressure who received streptokinase. Indeed, the point estimate of the difference in mortality between patients in this blood pressure strata who received alteplase rather than streptokinase was as great as that in any other cohort. This finding suggests that, contrary to previous recommendations [4], accelerated alteplase therapy is at least as beneficial as streptokinase in patients who present with acute myocardial infarction and very high blood pressure. The increased risk for stroke is outweighed by the decreased risk for death.

Study Limitations

The major limitation of our study is that we reviewed outcomes in relation to only one measurement of blood pressure: that obtained at the time of randomization. Blood pressure measurements obtained at other times may have equal or greater importance. Blood pressure measurements taken out of the hospital or after the administration of pain relievers or other medications may better reflect true values. In fact, the blood pressure measurements taken at study entry in the current study may have been substantially affected by medications given at the time of presentation. In addition, sustained hypertension after thrombolytic treatment may have a greater relation to cerebral hemorrhage than does a single abnormal value before treatment. Because of concern about the risks of administering thrombolytic therapy to patients who present with hypertension, especially given the caution communicated to GUSTO-I investigators at the time of enrollment, patients with very high blood pressures who were entered into the GUSTO trial may be a lower-risk sample that is not representative of the general population. This could limit the generalizability of our results to the overall population of hypertensive persons who have acute myocardial infarction.

Conclusions

As systolic blood pressures at study entry increased within the range of 100 mm Hg to about 175 mm Hg, the risk for death decreased but the risk for stroke increased. At systolic blood pressures greater than 175 mm Hg, the risk for stroke was markedly higher, more so in patients who received accelerated alteplase therapy. Despite the increased risk for stroke, however, the incidence of the combined end point of death plus disabling stroke was lower with accelerated alteplase treatment. In a small subgroup of patients with low risk for death from cardiac causes and high blood pressure at study entry, the risk for stroke may outweigh the survival benefit produced by thrombolytic therapy. More work must be done to determine the risk-to-benefit ratio of thrombolysis among patients with low risk for death from cardiac causes who present with hypertension. Whether decreasing blood pressure to less than 175 mm Hg reduces the risk for stroke without worsening mortality is also worth further investigation.

From Flinders Medical Centre, Bedford Park, Australia; Queens Medical Centre, Nottingham, United Kingdom; Hospital Beaumont, Dublin, Ireland; Green Lane Hospital, Auckland, New Zealand; Duke University Medical Center, Durham, North Carolina; and the Cleveland Clinic Foundation, Cleveland, Ohio.

Dr. Wilcox: Department of Medicine, Queen's Medical Centre, Nottingham NG7 2UH, United Kingdom.

Dr. Horgan: Beaumont Hospital, PO Box 1297, Beaumont Road, Dublin 9, Ireland.

Dr. White: Cardiology Department, Green Lane Hospital, Epsom, Auckland 1003, New Zealand.

Dr. Granger: Box 3850, Duke University Medical Center, Durham, NC 27710.

Dr. Califf: Box 31123, Duke University Medical Center, Durham, NC 27710.

Dr. Topol: Department of Cardiology, Cleveland Clinic Foundation, Desk F25, 9500 Euclid Avenue, Cleveland, OH 44195.